Controller XML Configuration

The following root XML tags are defined for all controller types:

Root XML Tag	Description
output	Parameters for simulation outputs generated by controllers within a simulation run. This includes debugging logs as well as gather metrics.
strategy	Parameters for different behaviors that robots can execute.
perception	Parameters for discretization of the continuous world into a grid, and/or the objects it tracks within it.
task_executive	Parameters for the task executive (the entitity responsible for managing/running tasks after they have been allocated), which are mainly configuration for HOW tasks are run once they are allocated.
task_alloc	Parameters for task allocation.
sensing_subsystem	Parameters for robot sensing subsystem.
actuation_subsystem	Parameters for robot actuation subsystem.
flocking	Parameters for swarms performing flocking.

output

• Required by: All controllers.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<output
    output_root="output"
    output_dir="__current_date__"
/>

• output_root - The root output directory in which the directories of different simulation runs will be placed. The path specified can be relative or absolute, and will be created if it does not exist.

• output_dir - The output directory for the current simulation under output_root. If you put the special field __current_date__ here, the simulation will get a unique output directory in the form YYYY-MM-DD:HH-MM.

strategy

• Required by: all.

• Required child attributes if present: [ nest_acq, nest_exit ].

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: [ blocks ].

XML configuration:

  <strategy>
    <nest_acq>
      ...
    </nest_acq>
    <nest_exit>
      ...
    </nest_exit>
    <blocks>
      ...
    </blocks>
/>

strategy/explore

• Required child attributes if present: [ strategy ].

• Required child tags if present: [ min_duration ].

• Optional child tags: none.

• Optional child attributes: none.

XML configuration:

  <strategy>
    ...
    <explore
       strategy="crw"
       min_duration="INT">
      ...
    </explore>
    ...
/>

• strategy - The strategy robots should use when exploring for SOMETHING (unspecified at this level). Valid values are:

  • CRW - Correlated Random Walk.

• min_duration - How long to perform the selected strategy for. Valid for: [ CRW ].

strategy/nest

• Required child attributes if present: none.

• Required child tags if present: [ acq, exit ].

• Optional child tags: none.

• Optional child attributes: none.

XML configuration:

  <strategy>
    ...
    <nest>
      <acq>
        ..
      </acq>
      <exit>
        ..
      </exit>
    </nest>
    ...
/>

strategy/nest/acq

• Required by: All controllers.

• Required child attributes if present: strategy.

• Required child tags if present: None.

• Optional child attributes: None.

• Optional child tags: None.

XML configuration:

<nest>
    <acq
       strategy="wander|random_thresh|wander_random_thresh"
       duration="INT"/>
    ...
</nest>

• strategy - The strategy robots should use once they have entered the nest with an object to choose a precise location to drop it at. Valid values are:

  • wander - Perform phototaxis+wander, avoiding collisions as needed, for a random number of timesteps before acquiring the nest.

  • random_thresh - Perform phototaxis+collision avoidance, choosing a random point along the vector pointing from where the robot enters the nest to the center to treat as the center/use as the nest acquisition point, and phototaxis to that point.

  • wander_random_thresh - random_thresh + wander.

• duration - How long to perform the strategy for. Valid for: [ wander ].

strategy/nest/exit

• Required by: All controllers.

• Required child attributes if present: strategy.

• Required child tags if present: None.

• Optional child attributes: None.

• Optional child tags: None.

XML configuration:

<nest>
    <acq strategy="wander|anti_phototaxis"/>
    ...
</nest>

• strategy - The strategy robots should use once they have entered the nest with an object to choose a precise location to drop it at. Valid values are:

  • wander - Perform wander, avoiding collisions as needed, until the robot exits the nest or gets pushed out.

  • anti_phototaxis - Perform anti-phototaxis+collision avoidance until the robot exits the nest.

strategy/blocks

• Required child attributes if present: none.

• Required child tags if present: none.

• Optional child tags: [ drop ].

• Optional child attributes: none.

XML configuration:

  <strategy>
    ...
    <blocks>
      <drop>
        ..
      </drop>
    </blocks>
    ...
/>

strategy/blocks/drop

• Required child attributes if present: [ strategy ].

• Required child tags if present: none.

• Optional child tags: [ duration ].

• Optional child attributes: none.

XML configuration:

  <blocks>
    ...
    <drop
       strategy="backup|backup_pivot"
       duration="INT">
      ...
    </drop>
    ...
/>

• strategy - The strategy robots should use to drop blocks. Valid values are:

  • backup - Backup for duration timesteps. Useful when robots are pushing blocks to “carry” them.

  • backup_pivot - Backup for duration/2 timesteps, then pivot for duration/2 timesteps. Useful when robots are pushing blocks to “carry” them.

• duration - How long to perform the selected strategy for. Valid for: [ backup, backup_pivot ].

perception

• Required child attributes if present: [ type ].

• Required child tags if present: none.

• Optional child tags: [ rlos ]

• Optional child attributes: none.

XML configuration:

<perception
  type="STRING">
  <rlos>
     ...
  </rlos>
</perception>

• type - The perception type to use.

perception/rlos

• Required child attributes if present: [ los_dim ].

• Required child tags if present: none.

• Optional child tags: [ arena ].

• Optional child attributes: none.

XML configuration:

<perception>
  ...
  <rlos
    dim="FLOAT">
    <grid2D>
      ...
    </grid2D>
  </rlos>
  ...
</perception>

• dim - The dimension of robot LOS (LOS is a square).

• grid2D - The dimensions of the arena (optional).

perception/rlos/grid2D

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<rlos>
  ...
  <grid2D resolution="FLOAT"
          dims="FLOAT, FLOAT, FLOAT"/>
  ...
</rlos>

• resolution - The size of the cells the arena is broken up (discretized) into. Should probably be the same as whatever the block size is, to make things easy.

• dims - The size of the arena, specified as “X, Y, Z” (the spaces are mandatory).

task_executive

• Required by: None. Used by all task based controllers with the default values shown below if it is omitted.

• Required child attributes if present: none.

• Required child tags if present: none.

• Optional child attributes: [update_exec_ests, update_interface_ests ]

• Optional child tags: none.

XML configuration:

<task_executive
    update_exec_ests="false"
    update_interface_ests="false"/>

• update_exec_ests - If true, then the executive will use the elapsed time since a task started to update the task time estimate. Estimate is updated on both abort an completion. Default if omitted: false.

• update_interface_ests - If true, then the executive will use the calculated interface time for a task to update the interface estimate for the task. Estimate is updated on both abort and completion. Default if omitted: false.

task_alloc

• Required by: None. Used by all task based controllers with the default values shown below if it is omitted.

• Required child attributes if present: all.

• Required child tags if present: task_abort.

• Optional child attributes: none.

• Optional child tags: [stoch_nbhd1, task_exec_estimates, epsilon_greedy ].

XML configuration:

<task_alloc
    policy="random|epsilon_greedy|strict_greedy|stoch_nbhd1|ucb1">
    <matroid_stoch_nbhd>
    ...
    </matroid_stoch_nbhd>
    <task_exec_estimates>
     ...
     </task_exec_estimates>
    <task_abort>
     ...
     </task_abort>
</task_alloc>

• policy - When performing task allocation, how should tasks be selected?

  • random - Choose a random task each time.

  • epsilon_greedy - Choose the greedy best with probability 1 - epsilon, otherwise choose a random task. Has provably bounds on regret, treating task allocation as a multi-armed bandit problem.

  • strict_greedy - A pure greedy matroid optimization approach.

  • stoch_nbhd1 - A stochastic greedy approach within the neighborhood of the most recently executed task (max distance is 1).

  • UCB1 - A deterministic greedy approach based on regret minimization (has provable logarithmic bound).

Many child tags in <task_alloc> use sigmoid-based functions for choosing between alternatives, with the input src and sigmoid method varying. For such tags, all child attributes and tags are required unless specified otherwise.

XML configuration:

...
<src_sigmoid_sel
    input_src="exec|interface">
    <sigmoid_sel
    method="harwell2018">
        <sigmoid reactivity="FLOAT"
                 offset="FLOAT"
                 gamma="FLOAT"/>
    <sigmoid_sel/>
</src_sigmoid_sel>
...

• input_src - Can be exec or interface, indicating that estimates of execution/interface times should be used in the selection process.

• method - The method used to calculate a probability using the selected input source.

• reactivty - Once the offset is tripped, this parameter controls how fast the probability a robot aborts its current task grows.

• offset - A positive proportition indicating what ratio of measured execution time to the robot’s best estimate of the actual execution time of the task is considered to be the threshold for a task taking too long, and should be aborted.

• gamma - A scaling factor that is applied to the overall calculated probability.

task_alloc/task_abort

Parameters governing task abort calculations.

• Required by: All task based controllers.

• Required child attributes if present: none.

• Required child tags if present: src_sigmoid_sel.

• Optional child attributes: none.

• Optional child tags: none.

method=harwell2018 is required.

XML configuration:

<task_alloc>
    ...
    <task_abort>
        <src_sigmoid_sel
            input_src="exec|interface">
            <sigmoid_sel
                method="harwell2018">
                <sigmoid reactivity="FLOAT"
                         offset="FLOAT"
                         gamma="FLOAT"/>
            <sigmoid_sel/>
        </src_sigmoid_sel>
    </task_abort>
    ...
</task_alloc>

task_alloc/task_exec_estimates

Parameters governing how/if task execution cost estimates (time costs) will be seeded and updated.

• Required by: None.

• Required child attributes if present: None.

• Required child tags if present: ema (only if seed_enabled is true).

• Optional child attributes: all. Only the task names used by the loaded task decomposition graph are required; others are ignored.

• Optional child tags: none.

XML configuration:

<task_alloc>
    ...
    <task_exec_estimates
        seed_enabled="false"
        task_name="2000:4000">
            <ema alpha="FLOAT"/>
    </task_exec_estimates>
    ...
</task_alloc>

• seed_enabled - If true, then all estimates of task execution times are initialized randomly within the specified ranges, rather than with zero, in order to avoid any possibly weird behavior on system startup. Default if omitted: false.

• <task name> - Takes a pair like so: 100:200 specifying the range of the uniform random distribution over which a robots’ initial estimation of the duration of the specified task will be drawn. Only used if seed_enabled is true. Valid values for <task_name> are project-specific.

task_alloc/task_exec_estimates/ema

The exponential moving average (EMA) parameters for task execution estimate updating.

• Required by: None.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<task_exec_estimates>
    ...
    <ema alpha="FLOAT"/>
    ...
</task_exec_estimates>

• alpha - Parameter for exponential weighting of a moving time estimate of the true execution/interface time of a task. Must be < 1.0.

task_alloc/epsilon_greedy

Parameters for the epsilon-greedy method described in :xref:`Auer2002`.

• Required by: Task based controllers if the selected policy is epsilon_greedy.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<task_alloc>
    ...
    <epsilon_greedy
        epsilon="FLOAT"
        regret_bound="log|linear"/>
    ...
</task_alloc>

• epsilon - Used to control exploration of the method. Must be between 0.0 and 1.0.

• regret_bound - What is the provable bound on regret?

  • log - Logarithmic bounded.

  • linear - Linearly bounded (more regret).

task_alloc/ucb1

Parameters for the UCB1 method described in :xref:`Auer2002`.

• Required by: Task based controllers if the selected policy is ucb1.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<task_alloc>
    ...
    <ucb1 gamma="FLOAT"/>
    ...
</task_alloc>

• gamma - Weighting factor to control how much exploration of the method. Must be between 0.0 and 1.0.

task_alloc/stoch_nbhd1

Parameters for the stochastic task allocation neighborhood method described in :xref:`Harwell2020a-demystify`.

• Required by: Task based controllers if the selected policy is stoch_nbhd1.

• Required child attributes if present: all.

• Required child tags if present: [ task_partition, subtask_sel ].

• Optional child attributes: none.

• Optional child tags: [ tab_sel ].

XML configuration:

<task_alloc>
    ...
    <stoch_nbhd1
        tab_init_policy="root|max_depth|random">
            <task_partition>
            ...
            </task_partition>
            <subtask_sel>
            ...
            </subtask_sel>
            <tab_sel>
            ...
            </tab_sel>
    </stoch_nbhd1>
    ...
</task_alloc>

• tab_init_policy - When performing initial task allocation, how should the initial Task Allocation Block (TAB), consisting of a root has and two sequentially interdependent subtasks, be selected. Valid values are:

  • root - Use the root TAB as the initially active TAB.

  • random - Choose a random TAB as the initially active TAB.

  • max_depth - Choose a random TAB from among those at the greatest depth within the task decomposition graph that is passed to the executive.

task_alloc/stoch_nbhd1/task_partition

Task partitioning parameters for the STOCH-NBHD1 method. Based on the work in :xref:`Pini2011`.

• Required by: Task based controllers.

• Required child attributes if present: none.

• Required child tags if present: src_sigmoid_sel.

• Optional child attributes: [always_partition, never_partition ].

• Optional child tags: none.

XML configuration:

<stoch_nbhd1>
    ...
    <task_partition
        always_partition="false"
        never_partition="false">
        <src_sigmoid_sel
            input_src="exec|interface">
            <sigmoid_sel
                method="pini2011">
                <sigmoid reactivity="FLOAT"
                         offset="FLOAT"
                         gamma="FLOAT"/>
            <sigmoid_sel/>
        </src_sigmoid_sel>
    </task_partition>
</stoch_nbhd1>

• always_partition - If true, then robots will always choose to partition a task, given the chance. Default if omitted: false.

• never_partition - If true, then robots will never choose to partition a task, given the chance. Default if omitted: false.

method tag can be one of [ pini2011 ] for performing the stochastic partitioning decision. Calculated once upon each task allocation, after the previous task is finished or aborted.

task_alloc/stoch_nbhd1/subtask_sel

Task partitioning parameters for the STOCH-NBHD1 method. Based on the work in :xref:`Pini2011`, :xref:`Brutschy2014`.

• Required by: Task based controllers.

• Required child attributes if present: none.

• Required child tags if present: src_sigmoid_sel.

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<stoch_nbhd1>
    ...
    <subtask_sel>
        <src_sigmoid_sel
            input_src="exec|interface">
            <sigmoid_sel
                method="harwell2018|random">
                <sigmoid reactivity="FLOAT"
                         offset="FLOAT"
                         gamma="FLOAT"/>
            <sigmoid_sel/>
        </src_sigmoid_sel>
    </subtask_sel>
    ...
</stoch_nbhd1>

method tag can be one of [harwell2018, random ] to perform stochastic subtask selection if partitioning is employed.

task_alloc/stoch_nbhd1/tab_sel

Parameters for Task Allocation Block (TAB) selection in recursive task decomposition graphs used in the STOCH-NBHD1 method described in :xref:`Harwell2020a-demystify`.

• Required child attributes if present: src_sigmoid_sel.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<stoch_nbhd1>
    ...
    <tab_sel>
        <src_sigmoid_sel
            input_src="exec|interface">
            <sigmoid_sel
                method="harwell2019">
                <sigmoid reactivity="FLOAT"
                         offset="FLOAT"
                         gamma="FLOAT"/>
            <sigmoid_sel/>
        </src_sigmoid_sel>
    </tab_sel>
    ...
</stoch_nbhd1>

method tag that can be one of [ harwell2019 ].

sensing_subsystem

• Required by: All controllers.

• Required child attributes if present: none.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: [ proximity_sensor, env_sensor ].

XML configuration:

<sensing_subsystem>
    <proximity_sensor>
    ...
    </proximity_sensor>
    <env_sensor>
    ...
    </env_sensor>
</sensing_subsystem>

sensing_subsystem/proximity_sensor

Parameters for proximity sensor configuration.

• Required by: All controllers.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: [ exp_decay ].

• Optional child tags: none.

<sensing_subsystem>
    ...
    <proximity_sensor
        fov="FLOAT:FLOAT"
        delta="FLOAT"
        exp_decay="BOOL"/>
    ...
</sensing_subsystem>

• fov - The angle range to the left/right of center in which obstacles are not ignored (outside of this range they are ignored, assuming the robot will be able to drive by them). Takes a pair like so: -1:1. Specified in radians.

• delta - Tripping threshold for calculations for obstacle detection. Dependent on the value of exp_decay.

• exp_decay - If true, then apply \(e^{-x}\) to all distance values before returning them to the robot controller. Only used if COSM is built for ROS; for compatibility of ROS robot controller code with ARGoS robot controller code, which does this by default.

sensing_subsystem/ground_sensor

Parameters for ground sensor configuration.

• Required by: All controllers.

• Required child attributes if present: none.

• Required child tags if present: [ nest, block, cache ].

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<sensing_subsystem>
    ...
    <ground_sensor>
      <nest range="FLOAT:FLOAT"
            consensus="INTEGER"/>
      <block range="FLOAT:FLOAT"
            consensus="INTEGER"/>
      <cache range="FLOAT:FLOAT"
            consensus="INTEGER"/>
    </ground_sensor>
    ...
</sensing_subsystem>

For each of [nest, block, cache], the following child attributes are required:

• range - The range of ground sensor values used to detect the object. Should be unique among all the types of objects to detect.

• consensus - How many of the ground sensors must have readings within the specified range in order for a detection to be triggered.

actuation_subsystem

• Required by: All controllers.

• Required child attributes if present: none.

• Required child tags if present: [ apf_manager, diff_drive ]

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<actuation_subsystem>
    <apf_manager>
    ...
    </apf_manager>
    <diff_drive>
    ...
    </diff_drive>
</actuation_subsystem>

actuation_subsystem/apf_manager

Parameters for the virtual forces used to control robot movement, based on the original paper :xref:`Arkin1987` and the tutorial in :xref:`SteeringTutorial`.

• Required by: All controllers.

• Required child attributes if present: none.

• Required child tags: [ nav ].

• Optional child attributes: none.

• Optional child tags if present: [ flocking ].

XML configuration:

<actuation_subsystem>
    ...
    <apf_manager>
      <nav>
        ...
      </nav>
      <flocking>
        ...
      </flocking>
    </apf_manager>
    ...
</actuation_subsystem>

actuation_subsystem/apf_manager/nav

Parameters for the virtual forces used to control robot movement, based on the original paper :xref:`Arkin1987` and the tutorial in :xref:`SteeringTutorial`.

• Required by: All controllers.

• Required child attributes if present: none.

• Required child tags: none.

• Optional child attributes: none.

• Optional child tags if present: [ avoidance_force, arrival_force, wander_force, phototaxis_force, path_following_force ].

XML configuration:

<apf_manager>
    ...
    <nav>
      <avoidance_force>
        ...
      </avoidance_force>
      <arrival_force>
        ...
      </arrival_force>
      <wander_force>
        ...
      </wander_force>
      <phototaxis_force>
        ...
      </phototaxis_force>
      <path_following_force>
        ...
      </path_following_force>
    </nav>
    ...
</apf_manager>

actuation_subsystem/apf_manager/nav/avoidance_force

The force which repels robots from other nearby robots and obstacles (robots do not distinguish between these two cases).

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<nav>
  ...
  <avoidance_force lookahead="FLOAT"
                   max="FLOAT"/>
  ...
</nav>

• lookahead - How far ahead of the robot to look for obstacles. Currently unused, but may be used in the future.

• max - Max value for the force, in m/s.

actuation_subsystem/apf_manager/nav/arrival_force

The force which attracts robots towards a goal and gets them to it.

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<nav>
  ...
  <arrival_force slowing_radius="FLOAT"
                 slowing_speed_min="FLOAT"
                 max="FLOAT"/>
  ...
</nav>

• slowing_radius - Radius around target inside which robots will slow down linearly to not overshoot their target.

• slowing_speed_min - The minimum speed robotics will linearly ramp down to. Should be > 0.

• max - Max value for the force, in m/s.

actuation_subsystem/apf_manager/nav/wander_force

The force which causes robots to wander randomly in the environment in a correlated random walk.

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: [ bias_angle ].

• Optional child attributes: none.

• Optional child tags: none.

<nav>
  ...
  <wander_force circle_distance="FLOAT"
                circle_radius_min="FLOAT"
                max="FLOAT"
                interval="INTEGER">
     <bias_angle>
     ...
     </bias_angle>
  </wander_force>
  ...
</nav>

• circle_distance - Scaling factor for force; applied to current velocity.

• circle_radius - Displacement (i.e. wander) circle radius; placed at circle_distance from the robot.

• max - Max value for the force, in m/s.

• interval - How many timesteps to skip between applying the force.

actuation_subsystem/apf_manager/nav/wander_force/bias_angle

Configuration for how the wander/bias angle should be calculated if the wander force is employed.

• Required by: none.

• Required child attributes if present: [ src ].

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<wander_force>
  ...
  <bias_angle src="normal|uniform|custom"
                max_delta="FLOAT"
                angles="FLOAT,FLOAT,FLOAT,..."/>
  ...
</wander_force>

• src - The bias angle distribution source.

  • normal - Use a normal distribution, truncated at 3 std deviations from 0 and wrapped into [-max_delta, max_delta ].

  • uniform - Use a uniform distribution [-max_delta, max_delta ].

  • custom - Specify a custom distribution via a list of specific angles which will be drawn from uniformly.

• max_delta - +/- Maximum amount of heading change for the wander angle (a random value is chosen in this range). Specified in radians. Only affects normal and uniform sources.

• angles - A comma separated list of angles specified in radians defining the custom bias angle distribution to draw from.

actuation_subsystem/apf_manager/nav/phototaxis_force

The force which attracts/repels robots towards/away from light sources.

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<nav>
  ...
  <phototaxis_force max="FLOAT"/>
  ...
</nav>

• max - Max value for the force, in m/s.

actuation_subsystem/apf_manager/nav/path_following_force

The force which guides robots along a specified path.

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<nav>
  ...
  <arrival_force slowing_radius="FLOAT"
                 slowing_speed_min="FLOAT"
                 max="FLOAT"/>
  ...
</nav>

• max - Max value for the force, in m/s.

• radius - Radius around each point along the path to consider at part of the point; i.e., reaching any point inside the radius is equivalent to reaching the exact location of the point.

actuation_subsystem/apf_manager/flocking

Parameters for the virtual forces used to control robot movement during flocking. These are distinct from navigation forces which are forces that individual robots use to accomplish things–flocking forces relate to how robots interact with each other.

• Required by: All controllers.

• Required child attributes if present: none.

• Required child tags: none.

• Optional child attributes: none.

• Optional child tags if present: [ alignment_force, constant_speed_force ].

XML configuration:

<apf_manager>
    ...
    <flocking>
      <alignment_force>
        ...
      </alignment_force>
      <constant_speed_force>
        ...
      </constant_speed_force>
    </flocking>
    ...
</apf_manager>

actuation_subsystem/apf_manager/flocking/alignment_force

The force which aligns an agent’s trajectory with the average of one or more neighbors. Originally from :xref:`FLOCK:Bagarti2018-stochfov`.

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<flocking>
  ...
  <alignment_force max="FLOAT"/>
  ...
</flocking>

• max - Max value for the force, in m/s.

actuation_subsystem/apf_manager/flocking/constant_speed_force

The force which aligns an agent’s trajectory with the average of one or more neighbors. Originally from :xref:`FLOCK:Bagarti2018-stochfov`.

• Required by: none.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

<flocking>
  ...
  <constant_speed_force max="FLOAT"/>
  ...
</flocking>

• max - Max value for the force, in m/s.

actuation_subsystem/diff_drive

Parameters for robot differential drive.

• Required by: All controllers.

• Required child attributes if present: all.

• Required child tags if present: none.

• Optional child attributes: none.

• Optional child tags: none.

XML configuration:

<actuation_subsystem>
    ...
    <diff_drive soft_turn_max="FLOAT"
                max_speed="FLOAT"/>
    ...
</actuation_subsystem>

• soft_turn_max - If actuators are told to change to a heading within a difference greater than the one specified by this parameter to the current heading, a hard turn is executed (spin in place). Specified in radians.

• max_speed - The maximimum speed of the robot, in m/s.

flocking

• Required by: none.

• Required child attributes if present: [ strategy ].

• Required child tags if present: none.

• Optional child attributes: [ leader_sel_prob ].

• Optional child tags: [ stoch_fov ].

XML configuration:

<flocking
   strategy="stoch_fov"
   leader_sel_prob="FLOAT">

    <stoch_fov>
    ...
    </stoch_fov>

</flocking>

• strategy - The flocking strategy to employ. Options are:

  • stoch_fov - A Stochastic Field Of View (FOV) method from :xref:`FLOCK:Bagarti2018-stochfov`.

• leader_sel_prob - The probability that a given agent will be selected as a leader. Can be 0.

flocking/stoch_fov

Parameters for Stochastic Field Of View (FOV) flocking method from the :xref:`FLOCK:Bagarti2018-stochfov` paper.

• Required by: none.

• Required child attributes if present: [ theta_max, mean_interaction_dist ].

• Required child tags: none.

• Optional child attributes: none.

• Optional child tags if present: none.

XML configuration:

<flocking>
    ...
    <stoch_fov theta_max="FLOAT"
               mean_interaction_dist="FLOAT"/>

    ...
</flocking>

• theta_max - The bearing angle defining the wedge FOV in which an agent looks to find another agent to interact with.

• mean_interaction_dist - The distance which agents should try to be from those which they interact with. Steering forces increase/decrease to try to maintain this distance at all times.

flocking/trajectory

Parameters for specifying trajectories which the swarm should follow during flocking.

• Required by: none.

• Required child attributes if present: none.

• Required child tags: [ point ].

• Optional child attributes: [ loop ].

• Optional child tags if present: none.

XML configuration:

<flocking>
   ...
   <trajectory loop="BOOL">
      <point position="FLOAT,FLOAT,FLOAT"/>
      <point position="FLOAT,FLOAT,FLOAT"/>
      ...
   </trajectory>

   ...
</flocking>

• loop - Whether or not the trajectory is a loop and should be repeated indefinitely.

• point - Each point along the trajectory. At least 2 points must be specified.